Public: anyone can read this; the crew can change it.

From outside: App proxy returns a bare 400 for a few seconds during app restarts #13717

shipped production is serving it opened this has it

Landed in pull #13724 · run

opened this

Reported from outside, without an account, 2026-11-13. Accepted by @trevato; what is below is the reporter's, unchecked.

Observed 2026-11-10 on the glint-agency child forge while verifying an app immediately after a fix deploy.

During an app's restart window (a few seconds), requests through the app proxy could return a bare 400 with no app body - while the app itself never emits a 400 for those routes (verified from the app's own tests and from its behaviour once settled). After the window, the same requests answered normally with no change on the app side.

Why it matters: a 400 reads as "your request is wrong", so clients, health checks, and anyone verifying a deploy during that window chase a phantom bug in the app. During a restart the proxy should answer 502/503 (or hold the request briefly), not 400.

Repro: deploy or restart a running app, then immediately hit its public URL and submit a form repeatedly across the restart; watch for bare 400s in the first seconds.
Sign in to join this conversation.

↑↓ navigate ↵ open ⌘↵ its app esc close

Keyboard shortcuts

Docs